Detecting Prolonged Stress in Real Life Using Wearable Biosensors and Ecological Momentary Assessments: Naturalistic Experimental Study.

BACKGROUND: Increasing efforts toward the prevention of stress-related mental disorders have created a need for unobtrusive real-life monitoring of stress-related symptoms. Wearable devices have emerged as a possible solution to aid in this process, but their use in real-life stress detection has not been systematically investigated. OBJECTIVE: We aimed to determine the utility of ecological momentary assessments (EMA) and physiological arousal measured through wearable devices in detecting ecologically relevant stress states. METHODS: Using EMA combined with wearable biosensors for ecological physiological assessments (EPA), we investigated the impact of an ecological stressor (ie, a high-stakes examination week) on physiological arousal and affect compared to a control week without examinations in first-year medical and biomedical science students (51/83, 61.4% female). We first used generalized linear mixed-effects models with maximal fitting approaches to investigate the impact of examination periods on subjective stress exposure, mood, and physiological arousal. We then used machine learning models to investigate whether we could use EMA, wearable biosensors, or the combination of both to classify momentary data (ie, beeps) as belonging to examination or control weeks. We tested both individualized models using a leave-one-beep-out approach and group-based models using a leave-one-subject-out approach. RESULTS: During stressful high-stakes examination (versus control) weeks, participants reported increased negative affect and decreased positive affect. Intriguingly, physiological arousal decreased on average during the examination week. Time-resolved analyses revealed peaks in physiological arousal associated with both momentary self-reported stress exposure and self-reported positive affect. Mediation models revealed that the decreased physiological arousal in the examination week was mediated by lower positive affect during the same period. We then used machine learning to show that while individualized EMA outperformed EPA in its ability to classify beeps as originating from examinations or from control weeks (1603/4793, 33.45% and 1648/4565, 36.11% error rates, respectively), a combination of EMA and EPA yields optimal classification (1363/4565, 29.87% error rate). Finally, when comparing individualized models to group-based models, we found that the individualized models significantly outperformed the group-based models across all 3 inputs (EMA, EPA, and the combination). CONCLUSIONS: This study underscores the potential of wearable biosensors for stress-related mental health monitoring. However, it emphasizes the necessity of psychological context in interpreting physiological arousal captured by these devices, as arousal can be related to both positive and negative contexts. Moreover, our findings support a personalized approach in which momentary stress is optimally detected when referenced against an individual's own data.

Self-regulation of stress-related large-scale brain network balance using real-time fMRI neurofeedback.

It has recently been shown that acute stress affects the allocation of neural resources between large-scale brain networks, and the balance between the executive control network and the salience network in particular. Maladaptation of this dynamic resource reallocation process is thought to play a major role in stress-related psychopathology, suggesting that stress resilience may be determined by the retained ability to adaptively reallocate neural resources between these two networks. Actively training this ability could hence be a potentially promising way to increase resilience in individuals at risk for developing stress-related symptomatology. Using real-time functional Magnetic Resonance Imaging, the current study investigated whether individuals can learn to self-regulate stress-related large-scale network balance. Participants were engaged in a bidirectional and implicit real-time fMRI neurofeedback paradigm in which they were intermittently provided with a visual representation of the difference signal between the average activation of the salience and executive control networks, and tasked with attempting to self-regulate this signal. Our results show that, given feedback about their performance over three training sessions, participants were able to (1) learn strategies to differentially control the balance between SN and ECN activation on demand, as well as (2) successfully transfer this newly learned skill to a situation where they (a) did not receive any feedback anymore, and (b) were exposed to an acute stressor in form of the prospect of a mild electric stimulation. The current study hence constitutes an important first successful demonstration of neurofeedback training based on stress-related large-scale network balance - a novel approach that has the potential to train control over the central response to stressors in real-life and could build the foundation for future clinical interventions that aim at increasing resilience.